MOLECULAR PHYLOGENETICS AND EVOLUTION Vol. 1, No. 2, June, pp. 136-147, 1992

Variation and the Phylogenetic Utility of the Large Ribosomal Subunit of Mitochondrial DNA from the Order JAMES N. DERR,* SCOTT K. DAVIS,* JAMES B. WooLLEY,t AND ROBERT A. Wwmord

*Department of Sciences and tDepartment of Entomology, Texas A&M University, College Station, Texas 77843

Received February 4, 1992; revised July 6, 1992

provide pivotal information for investigating these Nucleotide sequence variation from a 573-bp region questions. Nevertheless, clearly supported and robust of the mitochondrial 16s rRNA gene was determined models of the phylogenetic relationships for major com- for representative hymenopteran taxa. An overall bias ponents of the Hymenoptera are lacking. in the distribution of A and T bases was observed from The most recent and comprehensive investigations all taxa; however, the terebrants (parasitoids) dis- regarding hymenopteran phylogeny are those of Ks- played significantly lower AT ratios as well as a higher nigsmann (1976, 1977, 1978a,b) and Rasnitsyn (1980, degree of strand asymmetry. Moreover, a strong posi- 1988). A number of other recent investigations have tive correlation was observed between relative AT rich- examined either specific superfamilies within the Hy- ness and sequence divergence, suggesting selection at menoptera (Brothers, 1975; Carpenter, 1986; Gibson, the nucleotide level for A and T bases as well as func- 1986) or various character systems across taxa (e.g., tionality. Overall sequence difference ranged from 2.3 Saini and Dhillon, 1980; Gibson, 1985; Darling, 1988; to 53.4%, with the maximum divergence between mem- Johnson, 1988; Whitfield et al., 1989). While these bers of the two Hymenopteran suborders. These data were used in a phylogenetic analysis to illustrate the studies have generally increased our understanding of utility and degree of resolution provided by this infor- the morphological basis for interpreting relationships, mation at various hierarchical levels within this taxo- there is nonetheless little congruence among research- nomically diverse order. Parsimony analysis revealed ers regarding higher classification. strong evidence for monophyly of the aculeates and the Composition of the Symphyta varies depending on terebrants. Most noteworthy was a strongly supported whether the Cephidae (Kanigsmann, 1977) or the Orus- clade containing the two terebrant superfamilies Icheu- sidae (Rasnitsyn, 1980) are removed and placed as the monoidea and Chalcidoidea. Conversely, high se- sister group to the . Even when both are re- quence divergence values resulted in instability at the tained in the Symphyta (the traditional arrangement), base of the tree and limited resolution at the higher there are problems in establishing a symphytan lin- taxonomic levels. Nevertheless, these results do identify eage based on shared, derived features relative to the those taxonomic levels for which 16s rRNA sequences Apocrita (Gibson, 1986; Gauld and Bolton, 1988; John- are phylogenetically informative. 8 1892 Academic PWIS, hc. son, 1988) and the group is now widely regarded as paraphyletic (Gauld and Bolton, 1988). More detailed resolution of the Symphyta is crucial for understand- ing the origins of the Apocrita and in particular for establishing character polarities within Apocrita using INTRODUCTION outgroup arguments. The Apocrita are generally interpreted as monophy- The Hymenoptera is one of the more diverse and letic largely on the basis of the fusion of the first ab thoroughly studied insect orders. Traditionally, this dominal segment to the thorax. It is usually subdivided order is divided into two suborders, Symphyta and into the Aculeata and the Terebrantes (Parasitical. Apocrita, that together contain some 110,000 described The aculeate Hymenoptera (, solitary and social species. The Symphyta is composed ofthe , horn- , and ) almost certainly form a monophyletic tails, and parasitic wood wasps, whereas the Apocrita group (Brothers, 1975; Carpenter, 1986); however, se- contains the majority of the parasitic Hymenoptera, rious questions remain regarding the monophyly of the the ants, solitary and social wasps, and bees. The evo- terebrants (“parasitoids”) due to the lack of demon- lution of parasitism and social behavior in this order strated synapomorphies. has generated considerable interest, and recently Car- The parasitoids belong to at least nine currently rec- penter (1989) proposed that phylogenetic methods may ognized superfamilies whose relationships and compo-

136 1055-7903/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved. mtDNA 16s I-RNA SEQUENCE VARIATION FROM HYMENOPTERA 137 sition are the subject of some debate (Konigsmann, DNA Amplification 1978a; Rasnitsyn, 1980, 1988; Naumann and Masner, 1985; Gibson, 1986; Gauld and Bolton, 1988). Of partic- At the time we initiated this study, few oligonucleo- ular interest here is the recent suggestion that the tide primers were available for amplified specific DNA Aculeata and the terebrant superfamily Ichneumo- sequences from insect taxa. We used an edited version noidea ( -t- Braconidae) are sister of the primers first designed by T. Kocher (personal groups (Rasnitsyn, 1988; Mason, unpublished obser- communication) to amplify a region of the 16s rRNA vation). gene. These primers were chosen because they pro- Two attributes of the order Hymenoptera make it duced consistent PCR products from a number of differ- particularly appealing for investigations using molec- ent hymenopteran taxa. Primer sequences are repre- ular techniques. First, although a number of morpho- sented by the light (L) or heavy (H) strand of mtDNA logical-based studies are available for various groups and their position relative to the complete Drosophila within this order, few provide well-supported phyloge- mtDNA sequence (Clary and Wolstenholme, 1985). netic hypotheses. For example, many of these studies Both primers were constructed with 5’ terminal EcoRI are not based on sound phylogenetic methods and even restriction enzyme sites plus one additional base to fa- when they are, reductional features (such as loss of cilitate plasmid cloning of the amplified products. The wing veins), which are commonplace throughout the sequences of these two primers are as follows: primer order, often confuse attempts to assess character ho- 16saEcoRI (L-12,883), 5’-GGAATTCCCTCCGG’ITTG- mology. In addition, lack of agreement on even funda- AACTCAGATC-3’; and primer 16sbEcoRI (H-13,398), mental relationships hinders the interpretation of 5’-GGAA’Il’CCCGCCTG’ITI’ATCAAAAACAT-3’. Sub- character polarity. Second, the very small size of many sequently, “nested” primers were developed based on of the parasitic wasps (many Cl mm) prevents a molec- our initial sequence in comparison with the published ular genetic evaluation due to the lack of sufficient 16s rRNA sequences from Drosophila (Clary and Wol- material. However, techniques such as polymerase stenholme, 1985), Aeoks (Hsu-Chen et al., 19841, and chain reaction (PCR) offer new and powerful tools that Apis (Vlasak et al., 1987). The sequences of these prim- promise to provide new information that will greatly ers are 16saW (L-12,943),5’-TAATAATCAAACATA- facilitate future systematic research on this group of GAGGTCG-3’; and 16sbW (H-13,332), 5’-GACTGTT- . CAAAGGTAGCATAAT-3’. In this study we use PCR to illuminate DNA se- Template DNAs were amplified in 50 or 100~p,l total quence variation from the large ribosomal subunit reaction volumes with 30-35 PCR cycles (Saiki et al., (16s rRNA) of the mitochondrial genome from mem- 1988). Most primer/template combinations were am- bers of six superfamilies representing both suborders plified using the thermal stable DNA polymerase Taq (Symphyta and Apocrita) of Hymenoptera. These data (International Biotechnologies, Inc.) although Vent are used in a phylogenetic analysis to illustrate the DNA polymerase (New England BioLabs) was also utility and degree of resolution provided by this infor- used. All amplifications were performed with a Perkin mation at various hierarchical levels within this taxo- Elmer-Cetus thermal cycler with the following proto- nomically diverse order. For comprehensive reviews of col: 35 cycles of DNA denaturing at 93°C for 30 s, the utility of mitochondrial DNA (mtDNA) for evolu- primer annealing at 42-50°C for 60 s, and primer ex- tionary studies see the work of Moritz et al. (1987) and tension at 72°C for 2 min. With some template/primer Harrison (1989). combinations the insertion of a 4-min step cycle be- tween annealing and extension increased the yield of the final amplification product. One explanation may METHODS AND MATERIALS be the rapid stabilization of partially mismatched primer/template complexes by the DNA polymerase as DNA Extraction and Purification the reaction slowly warms from the annealing temper- DNA was extracted from frozen ( - SO’C) or ethanol- ature to (42-50°C) to the extension temperature preserved specimens using the proteinase K digestion (72°C). protocol of Maniatis et al. (1982). Final DNA pellets Following PCR, the amplified products were sepa- were washed in ice-cold 70% ethanol, redissolved in rated by electrophoresis in 1.5% agarose minigels (Ma- sterile water, and stored at 4°C. From the small-bodied niatis et al., 1982). Gels were stained with ethidium parasitic species, DNA was isolated following whole- bromide and the DNA was visualized by fluorescence body digestion of 5-15 individuals from a single iso- under uv light. Amplified products were concentrated line. Head, thorax, or larva provided the tissue source using Centricon 30 microconcentrators (Amicon Div., from larger specimens. High-molecular-weight DNA of W.R. Grace) by washing with 2 ml of ddH,O through sufficient quality for PCR was recovered from both fro- three rounds of centrifugation (5000 rpm, 3000 rcf) zen and alcohol-preserved samples. Refer to the Appen- with a Sorvall (Du Pant) RC-5 high-speed centrifuge dix for details regarding the specimens examined. using a SA-600 rotor. Typically, 50 ~1 of PCR product 138 DERR ET AL.

was recovered after inversion of the Centricon 30 tube removing inferred gaps from the analysis prevents the and centrifugation at 750 rpm (80 rcf) for 3 min. inclusion of spurious information resulting from the Amplified products were cloned into plasmid vectors loss of one to a few nucleotide bases in these apparently using two different protocols. First, following concen- hypervariable regions. Initial analyses included align- tration, PCR products and supercoiled pUC plasmids ment with the published 16s rRNA sequences from were digested with EcoRI and then ligated to form cir- Drosophila (Clary and Wolstenholme, 1985) and Aedes cular hybrid DNA molecules, transformed into compe- (Hsu-Chen et al., 1984) for outgroup comparison. Boot- tent bacterial cells (GIBCO BRL), and screened for re- strap resampling procedures were performed with 100 combinant clones. Other PCR products were ligated, replicates to investigate the “robustness” of the data transformed, and grown using the TA Cloning kit (In- set, and tree length distributions were examined to vitrogen) following the manufacturer’s directions. Al- test for “skewness” using the g, statistic of Sokal and though recent studies document a low error rate Rohlf (1981). This statistic is a common measure of frequency for Tuq polymerase in PCR (Eckert and skewness and is negative for distributions with left- Kunkel, 1990; Kwiatowski et ul., 1991), we sequenced skew, 0 for symmetric distributions, and positive for two to six clones from each individual to ensure poly- distributions with right-skew (Hillis and Huelsenbeck, merase fidelity. No PCR-based errors were observed. 1992). As an additional test of polymerase accuracy, we used purified double-stranded PCR products in combination RESULTS AND DISCUSSION with unbalanced priming to generate single-stranded Sequence Variation product for direct sequencing from selected individuals following the methods of Allard et al., (1991). All se- Multiple sequence alignment of the amplified region quencing reactions were accomplished using Seque- from selected Hymenoptera taxa to published homolo- nase 2 (U.S. Biochemical) following the Sanger method gous sequences from Aedes and Drosophila resulted in @anger et al., 1977). Priming for template DNAs was a matrix consisting of 573 positions (Fig. 1) due to nu- accomplished using the forward or reverse M-13 se- merous inferred deletion/insertion events (Table 1). quencing primers or one of the 16s rRNA PCR prim- Fragment sizes from individual taxa ranged from 510 ers. Both strands were sequenced for all template bp (Apis) to 552 bp (Allorhogus). Multiple individuals, DNAs in order to obtain sequence information close to each with identical nucleotide sequence, were analyzed the primer and to verify internal sequence. from the genera Allorhogus (2 ), (2 ani- mals), Xanthopimpla (2 animals), and Aphytis lingnu- Sequence Alignment and Analysis nensis (10 animals from a single isoline). The 2 Digo- Sequence alignment was accomplished using the nogustru individuals examined differed by one CLUSTAL package for multiple sequence alignments nucleotide substitution at position 184 (T-C). Two spe- of Higgins and Sharp (1988, 1989) and by hand. A cies, Polistes versicolor and the isoline of Aphytis yano- number of sequence alignments using the subroutines nensis, were analyzed by cloning the double-stranded CLUSTALl, CLUSTALB, and CLUSTAL4 with differ- PCR fragments and by direct sequencing of single- ent penalties for inferred gap events were considered. stranded PCR products. No sequence differences were Aligned sequences of the study group and outgroup observed between the cloned and the single-stranded taxa were analyzed cladistically by the maximum par- PCR products. simony procedures available in PAUP 3.0 using an ex- Nucleotide sequence information to date suggests haustive search or the branch-and-bound algorithm that insect mtDNA is very A and T rich (Clary and (Swofford and Olsen, 1990; Swofford, 1985). The most Wolstenholme, 1985; Hsu-Chen et al., 1984; Vlasak reasonable sequence alignment, judged by simply eval- et al., 1987; Crozier et al., 1989; Simon et al., 1990). We uating the alignment by eye and comparing the length found an overall bias in the number of A and T bases and consistency of the final trees from the parsimony expressed from all taxa ranging from a high of 0.794 analyses, was that found using the following CLUS- in Tremex to a low of 0.533 in Digonogustra (Table 1). TAL parameters: a gap penalty equal to 15 and a pen- However, representatives of the superfamilies Ichneu- alty of 5 for each additional position inserted within a monoidea and Chalcidoidea (i.e., Xunthopimplu, Digo- gap. Because of the difficulties in coding and in de- nogastra, Allorhogas, and Aphytis) have significantly termining homology among inferred gap events, we less AT-rich 16s rRNA sequences than the remaining conservatively coded all inferred gaps as missing data. taxa based on a least significant difference test using Although some useful information may be disregarded, a confidence interval of 95% (Table 1). In addition, the

FIG. 1. Aligned nucleotide sequence from a 573-bp region of the 16s rRNA gene of selected hymenopteran taxa and three previously published sequences: Drosophila (Clary and Wolstenholme, 19851, Aedes (Hsu-Chen et al., 1984), and Apis (Vlasak et al., 1987). Sequences are shown in reference to the Drosophib sequence with nucleotide substitutions expressed as the new bases, inferred gap events as hyphens, and positions identical to the Drosophila sequence as periods. mtDNA 16s rRNA SEQUENCE VARIATION FROM HYMENOPTERA 139

60 --GTMGMT-TTAAAAGTCGAACAGACTTAAA--ATTTGT AT...... C...G...... C...... C...A.....TT...... T .. ..A Tenthredinidae AT....-...... G...... C...... T...A...G.TT..T....T...T. AT .. ..A.T...... G...... C....M..AAAT..--ATT.T....TTT.T. Bnia AT...... T...... MC...A...-.TA..G.G..T..TT. AT...... T...... MC...A.....TA..G.G..T...T. AC...G..C...... TC..T.....A .. ---G.ACC...A.TAG.....G....-.TCG G AC...G..C...... TC..T.....A..---G.ACC...A.TAG.....G....-.TC GG AC...G..C...... TC..T.....A..---G.ACC...A.TAG.....G....-.TCG G AC...G..C...... TC..T.....A..---G.ACC ... A-TAG...... -.TlT G 8. vanonensis AC...G..C .G.. .. TC ..T.. ... A..---G.ACC...A.TA G.. ... G.. .. -.l'TTG

120 TATATCTTAATCCAACATCGAG-GTCGCAATCTTTTTTATCGATATGAACTCTCCAAAAA ..AT...... G...... A...... M ..... Tenthredinidae .-...T...... TA...A.....T.A- ...... T..AT.....T...... A...... TA...A.....T..A ...... -C...... T .. ..G ...... A.A.C...... T.....GT..A..A...G. Polistea .-C...... T .. ..G ...... A.A.C...... T.....GT..A..A...G. G..G..CGT.G...... A....T..A.CC.A..G.T...... G.....AG..T .G G. .G. .C.G ...... T..A.CC.A..G.T...... G.....AG..T .G G. .G. .C.G ...... T..A.CC.A..G.T...... G.....AG..T .G A. linananansis GG.G..C.G...... T..A.C..C ..G ...... G.G..G..TG..G .. A. vanonansis GG.G..C.G...... T..A.C..A..T.T...... GG....TG..G ..

180 ProsODhih AATTACGCTGTTATCCCTAAAGTAACTTAATTTTTTTAATCATTATW,ATGGATCAATT-- As&s ...... G...... A..M.T...... A.- - Tanthredinidae ...... G....T...T.C.M....TT...AG..-A.....TA .TT ... ..T...... GG....T...-AC.C-....T-....ATT-A.....TA.T T T ...... G....T...T.C...... T.CA...T..MT....AAA T Poliste T ...... G-...T...T.C..A....T..A.AA...TATC.TTAAA G A G...G...... GG...... GT.CCG..GG...AGT-..T...... G A G...G...... GG...... GT.CCG..GG...AGT-..T...... G A QbOnOaaStra G...G...... GG...... GT.CCG..GG...AGT-..T...... G A A. limnanensis G...G...... GGG...... GG..CGA...... G.TA..C...G ...... TT A. yWKmenSiS G..GG...... GG...... GG..CM...... GCTG..C...G.....A TT

240 OrOSODhila ATTCATAAATT-AATGTTT~TTM-MTWLA&J,----GTTTTT------T~ A!des ..A .. ..T ...... AAACA...T.A...... M ...... Tenthredinidae ... ..AT...... T....A.MT....A.....M.....AAA ...... Tremex .A-..AG.TA...T..CA..A.T.ATCAA.T.TTATT...CA.AT A ...... &.& .. -.T.T.TA.C..AA..AAA.CT.T.--..T.M ...... A ...... pm T.-...T-TA.CT...... A...... --.TT.M.....C AA Xanthwiinrzb G.AT.GT.G..CGC.T.GAC.GGTG..G.CTT.GCAT..AC.GCTCGGAGG....T.GG G All rho- G.AT.GT.G..CGC.T.GAC.GGTG..G.CTT.GCAT..AC.GCTCGGAGG....T.GG G A G.AT.GT.G..CGC.T.GAC.GGTG..G.CTT.GCAT..AC.GCTCGGAGG....T.GG G A. llnananensis .GATT.G.C..GT....C..GGT.AG...TT- - ...... C..ATGCTCGGAAG..T.T .G A. vanonensis .----.GGC..GTG..GCC.GGGTAT..-G.GGG AA . ..C.CATGGTATTTGGAAG.TC. 300 mOSODhilg TTTTM-T---AT-CACCCCAA-TAAAATATTT-l’AATTTATTA-AA--ATTAA--ATTM Asas ...... C...... G...... T.....TAAA.C..T...... T ...... T Tenthredinidae .. ..TTC ...... G ...... A..TTAA..A-.T...... T....C.T. zcaxm!a ... ..TT.TAA..AT T ...... T.A.ATT.ACC...... TAT..T....A ..C && ... -.CCA ...... CT...... C .... T.AA.CT.A-.A..T.TTT...... -- .- &J&gg#.. . ..-..T.....C..CT ...... --AAA..T.A-.C.-T..TA...... -- ..C.GC.CCGAGGT.G...... CCG....T .. -A.TGCAGG.TTGGATG..T.GG.CCT G Worhoaas ..C.GC.CCGAGGT.G...... CCG .. ..T . ..A.TGCAGG.TTGGTAG..T.GG.CCT G Qiaonoaastz3 ..C.GC.CCGAGGT.G...... CCT....T .. -A.TGCAGG.TTGGTAG..T.GG.CCT G A. linananensis .. -.TT.CCTA.GT.G...... C.T...AT.....GCCA---TAGCT.-----AG.C.T G A. -ensis .. ..GTTCCMGGT.G...... CCG...G----..GC.A---TT.GT.-----GGG--- - 140 DERR ET AL.

360 DrosoDhila TCTTTATM-T--TAAAATATM-M--T-ATAAAGATGGGTCTTCTCGTCTT tie&s ... ..T...C.....T...T..-...... A...... C ...... Tenthradinidae .T.A.T...T.AA...... T...... TG..T.ATTC...... C. Ll!laQB .T.A...... -.TTT...... TATAA..T..ATTC.- ...... A && -T.A-TA..TAAA.-TT.A...... -.-.T..ATTC...... A.....c C polistes ATCACTA..TAAAA.TTT...... CA.TT.T..ATTC.-...... A ...... A s .GGG.T.GTTAGG..CTG.TTGC.TT..TAAAT.....C.CC ...... Allorhoa .GGG.T.GTTAGG..CTG.TTGC.TT..TAAAT.....C.CC ...... Dia0noaastx.a .GGG.T.GTTAGG..CTG.TTGC.TT..TAAAT...-.C.C C ...... A. linsnanensis .A.-. T..GT.GT.--TG.T------TAAAT.....C.CC.C ...... 8. ym...... ----.T...TAGG.--T------..AGT.T...C..C.C 420 TTMATTAATTTTAG----CTTTTTGACT AAAAAATAAAATTCTATTTTAAATTTMATG .. ..T.AT...... A...... A.T...... M- ...... M Tenthredinidae ... ..AA..A....A...G...... A...T...... T...... A...... T..A . ..AA .A.TT.A...... T.TTCA...... T..A....TT.T...... A.AAA.GTT.M..G .. BBis A...T.A..-...TA...GM....A...... T.T..T...A---..T...MTTTA. Polistes A ...... T.-..AAA...G.A....A...TT...T.T...... A.-A...... M.TTA. XanthoDimala GCTGT----G.CAT.CCCG.C.C..C..GOGC.CC.C..T...AC.GG.T..-MGT.A. m GCTGT----G...T.CCCG.C.C..C..GGGC.GG.C..T...AC.GG.T..-MGT.A. oastrg GCTGT----G...T.CCCG.C.CGGC..GGGC.GG.C..T...AC.GG.T..-MGT.A. 8. linananensis A.GGG----G...T.CCAG.....GT...GGC.G..C.GT...AC.GA.TTG- ...... T A. - A.ATT----G.CATTCTAG.....GT...GG..G..C.GT..MT.GA.GGG-...C.A. 480 AAACAGTTAATATT-TCGTCCAACCATTCATTCCAGCCTTCAAT-TAAAAGACTAATGAT .G.....AT....C...A.T...... A.A...... -...... C .... Tenthredinidae .G....MTT..C....A..A..T...... T...... T...-...... T.T .. .G.....ATT...... M...... A..TT..T...-.....A..A.T..G. Aas .G...... T...... A.TA.TT.T.....A.A.TT...... -...... A.T.T .. Polistas .--...A.T...... TA..A.TT.T...... A..TT...... C.....A.....A ... B .G .... --TGA.CCC....GG.G...... A.AG.T.CCT.-.T...GGA..A.GA ... Allorhoaas .G....-CTGA.CCC....GG.G...... A.AG.T.CCT.-.T...GGA..A.G .... piaonosastra .G....-CTGA.CCC....GG.G.....----.AG.T.CCT.-.T...GGA.TA.G .... A. linsnanensis .G.....A.-.C..CC..A.TT .G ...... A.TG.T.ACT.-.T...GG...A.G .... A. yanonensis .G...... TGG.C.C..A.GTGG..- ..... A.TG.T..A..-.T...T....A ...... 540 DrosoDhila TATGCTACCTTTGCACAGTCAAAATACTGCGGCCATTTAAGTGGGCAGG neder ...... c...... c .... .T..C.A..CCA ...... Tenthredinidae .T ...... T...AAT...TA.A .. ..A Tremex ...... A...... C...... A.T.C..C .. -----AAT...... A .... &As ...... T...... C...... A..T ...... --AAT...T..A .. ..A polistes .T...... T...... T...... A...T...... T.-.MT...A ...... A XanthoDimDla ...... T.GGG...C...... G..A..C..G.G.....C ...... Allorhouas ...... G..T.GGG...C...... G..A..C..G.G.....C ...... Diaonosastra ...... G..T.GGG...C...... G..A..C..G.G.....C ...... 4. lmsnan ens15 ..C ...... G..T.GG....C...... G..A..G-.G.T.....C ...... 9. yanonensis ..C ...... G..T.GGG...C...... G..A..G..A.T.C...C ...... 573 DrosoDhila TTAGACTTTATATA----TMTT-CAAAAAGAC Aedes ...... A..?....A...A ...... Tenthredinidae ..G...... A..T....A.T..T ...... Tremex CC.T.TA.ATA..T....A..A...T...TA .A && .CGC...AA.AT...... --T...C.G ... Polistes A..A....GTC..T...... C-T...C ..... XanthoDimola CGGTG.C.CTA ... CTGG.G..G..T.G.G.T G m CGGTG.C.CTA...CTGG.G..G..T.G.G ..T Diaonosastra CGGTG. CACTA ... CTGG.G..G..T.G.G ..G 8. 1 1 nqnanenSiS CG.C..C..TA...CTTG..T.G..T...G.C T 8. y a nonensis CATTG.C..T A . ..CTGG.TTGGG.T...G.T T FIG. l-Continued mtDNA 16s rRNA SEQUENCE VARIATION FROM HYMENOFTERA 141

TABLE 1 Sizes of DNA Fragments Amplified and Sequenced, A + T Ratio, and Strand Asymmetry Base A+T Strand pairs ratioa asymmetry* 0 loo zw 240 400 Drosophila yakuba 512 0.766 6.066 nucleotide position Aedes albopictus 515 0.755 0.146* Treme~~ columba Apis mellifera 510 0.776 0.078 Polistes versicolor 516 0.777 0.093 Tenthredinidae 525 0.787 0.061 I- 60 - Tremex columba 528 0.794 0.053 * Y Xanthopimpla stemmator 551 0.534 0.111 Alldwgas pyralophagus 552 0.536 0.123 Digonogastra kimballi 546 0.533 0.132 I& zia a& do Aphytis yanonensis 527 0.594 0.127 nucleotide position Aphytis lingnanensis 518 0.573 0.154 Pollstes verskolor n Significant difference (indicated with vertical bars) among A + 100 T ratios and among strand asymmetry values is calculated with a -.- 1 95% confidence interval. w- * Strand asymmetry calculated by the following formula: (abs(No. ‘; 60- --- *a0 A-No. T) + abs(No. G-No. C))/(No. A + No. T + No. G + No. Cl. ae Values range from 0.0 (No. purines = No. pyrimidines) to 1.0 (one 2. - _.-._““._““,.” ___.,.. ,__-” .-_-_.,...I.--... ------_-_.-.__-__- -..--.?“--..-.-.. strand all purines the other all pyrimidines) (Smith et al., 1983). " -I- * This value is not significantly different from that of the lower 100 &I 3Lk 400 nucleotide position group. Aphytis ltngnanensls 100 I I I I ratio of A and T bases is not consistent throughout this region of the 16s rRNA gene. For example, as shown by the four representatives depicted in Fig. 2, there are areas of the sequence that display a trend toward 2w 300 400 higher A and T ratios (roughly corresponding to posi- nucleotide position tions 175,280,350, and 500) and other areas relatively rich in G and C bases around positions 80 and 460. FIG. 2. Distribution of A and T bases throughout the 16s rRNA These trends are apparent across all taxa examined. region sequenced from four representative taxa. The sequence of the mosquito (Aedes albopictus) was previously reported by Clary and In order to examine the relationship of nucleotide Wolstenholme (1985) and is the only dipteran shown. Of the hyme- composition to sequence divergence, we calculated an nopteran taxa, Tremex columba is a representative of the suborder adjusted divergence (Adj. Div.) value for all pairwise Symphyta, whereas both the aculeate (Polistes versicolor) and the combinations of sequences using the equation terebrant (Aphytis lingnanensis) are in the suborder Apocrita. Val- ues were calculated using a 50-bp sliding window with the sequence analysis package available in MacVector (International Biotechnolo- 11 - 4(0.25)2] Adj. Div. = D. gies, Inc.). 1 - (PA2 + PT2 + PG2 + PC2)

The numerator represents the probability of a replace- approaches 50%. This is compared with the probability ment by a different base if the nucleotide ratio is bal- of replacement in sequences with balanced nucleotide anced. The denominator is the probability of a replace- ratios, i.e., approximately 25%. Therefore, comparisons ment by a different base given the observed nucleotide of sequences with a nucleotide bias will give lower per- ratio within a “window” of 25 to 50 bp. The values PA, centage divergence estimates. Although a number of PT, PG, and PC are the percentage of each nucleotide published statistics are available for estimating nucle- expressed from both sequences in the window. The otide substitution, for our purposes it was necessary to value D is the total number of nucleotide replacements limit sequence comparisons to user-defined and rela- divided by the total number of positions in the window. tively small “windows” that were sensitive to rate The advantage of using Adj. Div. values of this fashion changes over short regions. We found that 25- to 50-bp is explained by the following example. In comparisons windows were long enough to show trends in the data between two sequences that are highly A and T rich, but not so short as to be overly influenced by unequal the probability that a single base substitution will not inferred gap events. In addition, this equation does not result in a different nucleotide base at that position weight transversions over transitions because regions 142 DERR ET AL.

0 / I I

1 101 201 301 401 501 Window Position FIG. 3. The combined A and T ratios (upper line) from a region of the 16s rRNA gene compared with the adjusted divergence (A$ Div.) (lower line) values from Apis and Aphytis lingnunensis. See text for details regarding the calculation of plotted values. rich in A and T bases show an unusually high degree sponding areas of high sequence divergence. A test of of transversion mutations; i.e., any substitution (A-T this observation was completed by comparing nonover- or T-A) is a transversion mutation. lapping 25bp windows between pairs of sequences us- The graph in Fig. 3 was constructed by limiting our ing a correlation coefficient (r> (Sokal and Rohlf, 1981). analysis to homologous sequences 50 bp long (window Specifically, we tested to see if areas high in A and T size) and calculating the Adj. Div. between two se- bases were statistically correlated with higher degrees quences based on the observed ratio of nucleotides. of sequence divergence. Resulting T values for all pair- This 50-bp window was then moved down the sequence wise comparisons (Table 2) were >O (indicating a posi- 1 base and the divergence estimate was recalculated. tive correlation between divergence and A-T richness) Repetitive cycles of this procedure were completed for for 52 of 55 tests. These results imply selection for nu- the entire sequence for all pairs of taxa. Figure 3 shows cleotide sequence at two distinct levels. The first in- the Adj. Div. values and combined A + T ratios of two volves selection for functionality of the ribosomal RNA representative sequences (Apis and Aphytis lingnu- molecule. However, selection must also proceed at a nensis) plotted by window position. This analysis high- second level; in areas where the sequence appears less lighted multiple areas rich in A and T bases and corre- constrained by functional requirements, nucleotide re-

TABLE 2 Matrix of Correlation Coefficients (r) from Comparisons of Combined A and T Ratios to Nucleotide Divergence Estimates from a Region of the 168 rRNA Gene from Insect Species

1 2 3 4 5 6 7 8 9 10 11

1. Drosophila yakuba - 0.498 0.538 0.393 0.612 0.490 0.490 0.358 0.436 0.761 0.752 2. Aea!es albopictus - 0.466 0.476 0.558 0.452 0.466 0.371 0.436 0.692 0.781 3. Tenthredinidae - 0.372 0.750 0.588 0.390 0.283 0.350 0.665 0.768 4. Tremex columba - 0.607 0.423 0.424 0.381 0.415 0.563 0.683 5. Apis m&fern - 0.490 0.313 0.289 0.276 0.505 0.763 6. Polistes versicolor - 0.375 0.306 0.341 0.710 0.346 7. Xanthopimpla stemmator - - 0.424 - 0.356 0.135 0.289 3. Digonogastm kimballi - - 0.200 0.206 0.276 9. Allorhogas pyralophagus - 0.220 0.289 10. Aphytis yanonensis - 0.441 11. Aphytis lingnanensis -

Note. Critical t values: to.l = 1.72 r > 0.360, to.M = 2.08 r > 0.420, to.ol = 2.34 r > 0.537. mtDNA 16s rRNA SEQUENCE VARIATION FROM HYMENOFTERA 143 placement seems to involve a propensity for the incor- nucleotide positions with the potential to contribute poration of A and T bases. Accordingly, these areas phylogenetic information. would display higher rates of sequence divergence, due This 16s rRNA nucleotide sequence was used to esti- to the lack of functional constraints, as well as a rela- mate hymenopteran phylogeny by unweighted parsi- tively higher number of A and T bases. What remains mony. Initial analyses using sequence from the two unexplained is the strong bias for A and T bases among dipteran taxa as outgroups (Drosophila and Aedes) re- nonvertebrate mtDNA sequences. sulted in the production of a single most parsimonious Nucleotide composition among different sequences tree consisting of 752 steps and a consistency index may also be compared on the basis of the degree of (CI, excluding uninformative’ characters) of 0.711. strand asymmetry as expressed by the number of pu- However, this tree could not be rooted with the Hyme- rines and pyrimidines on each DNA strand. As de- noptera as monophyletic in comparison with two dip- picted in Table 1, taxa with significantly lower ratios teran outgroups. Given that Hymenoptera are cer- of A and T bases (i.e., the superfamilies Ichneumo- tainly monophyletic (e.g., see Gauld and Bolton, 1988), noidea and Chalcidoidea) also display significantly this may indicate that Diptera are too remote an out- higher degrees of strand asymmetry. Although the cor- group to be useful in polarizing these data. However, it relation between function and strand asymmetry is at is noteworthy that overall sequence divergence values present unclear (Smith et al., 1983), these findings may (Table 3) between Diptera and some ingroup taxa are well represent differences that are responding to fac- less than the divergence between some ingroup (Hyme- tors above the nucleotide position level. noptera) taxa. In addition, exhaustive examination of A nucleotide substitution matrix was constructed all topologies resulting from the use of two dipteran from all pair-wise combinations of sequences using Ki- outgroups revealed two trees of 753 steps, one of which mura’s (1980) two-parameter model with the REAP has Hymenoptera as monophyletic. A second analysis computer package of McElroy et al., (1991) (Table 3, limiting outgroup comparisons to sequence variation below diagonal). This algorithm computes distance from Aedes produced a single tree with 701 steps (CI = values based on the K estimates of Kimura (1980, Eq. 0.753) (Fig. 4a). A tree-length distribution containing (10); Nei, 1987, Eq. (5.5)) that represent an estimate trees from 701-1200 steps was significantly more of the expected number of nucleotide substitutions per skewed than expected from random (9i = - 0.76; P < site between any pair of sequences. For comparison, 0.01). Asymmetrical tree distributions of this nature the percentage sequence difference based on each pair- have been proposed to be more informative due to the wise grouping of sequences from Fig. 1 is also provided possibility of increased resolution among the near- (Table 3, above diagonal). For Hymenoptera taxa, per- optimal solutions (Hillis and Huelsenbeck, 1992). A centage sequence difference ranged from less than bootstrap majority-rule consensus tree (Fig. 4b) has 2.5% (d = 0.023, Digonogastra vs Allorhogas) to more a topology identical to that of the most parsimoni- than 50% (d = 0.534, Tremex vs Aphytis lingnunensis). ous tree. Two major groups of hymenopteran taxa emerged: the first includes both members of the subor- Phylogenetic Analysis and Implications der Symphyta (Tenthredinidae and Tremex) and the mtDNA Sequences. Two hundred (35%) of the 573 two aculeates; the second is composed entirely of the nucleotide positions were identical in all taxa exam- terebrants. In addition, all branches are well supported ined and 86 (15%) variable positions differed by a sin- in the bootstrap consensus tree except for the two in- gle substitution in one taxon. This left 287 (50%) ternodes below the Symphyta and Aculeates and below

TABLE 3 Percentage Sequence Difference (Above) and Kimura’s Evolutionary Distance (1980, Eq. (10)) Below 1 2 3 4 5 6 I 8 9 10 11 1. Drosophila yakuba - 0.143 0.209 0.329 0.302 0.298 0.462 0.452 0.467 0.420 0.470 2. Aedes albopictus 0.164 - 0.206 0.322 0.307 0.309 0.452 0.441 0.459 0.422 0.482 3. Tenthredinidae 0.224 0.225 - 0.293 0.268 0.273 0.483 0.473 0.489 0.453 0.495 4. Tremex columba 0.363 0.346 0.307 - 0.352 0.356 0.511 0.497 0.513 0.499 0.534 5. Apis m&fern 0.325 0.342 0.295 0.390 - 0.185 0.497 0.488 0.500 0.503 0.497 6. Polistes versicolor 0.311 0.337 0.301 0.400 0.185 - 0.507 0.499 0.513 0.511 0.520 7. Xanthopimpla stemmator 0.576 0.566 0.649 0.747 0.648 0.693 - 0.024 0.036 0.284 0.331 8. Digonogastm kimballi 0.582 0.577 0.662 0.754 0.648 0.712 0.044 - 0.023 0.265 0.323 9. Allodwgas pymlophagus 0.560 0.552 0.637 0.123 0.629 0.689 0.037 0.028 - 0.281 0.335 10. Aphytis yanonensis 0.449 0.464 0.529 0.640 0.611 0.642 0.295 0.291 0.278 - 0.216 11. Aphytis lingnanensis 0.498 0.545 0.596 0.689 0.604 0.653 0.336 0.336 0.323 0.201 - 144 DERR ET AL.

a w-asWI redinidae, Tremex, the Aculeata, and the Terebrantes Aedes form a basal multifurcation in the consensus tree, but 3343 w I I Tenthredlnidae the sister group relationship between Chalcidoidea and is retained. Restricting the parsimony analysis to 16s rRNA se- quence variation from Hymenoptera and using either of the symphytan sequences as an outgroup resulted in a single tree with 639 steps and a CI of 0.79 (Fig. 5a, shown with Tenthredinidae as the outgroup). As Allorhogao before, the aculeates and terebrants form distinct mo- nophyletic groups that are reconstructed in 100% of the bootstrap replications (Fig. 5b) and the tree-length distribution is significantly skewed to the left @I = - 0.76; P < 0.01). Among the terebrants, the three rep- b resentatives of the superfamily Ichneumonoidea also I Tenthredlnidae consistently cluster together. Sequences from this re- gion of the 16s rRNA gene differ at 19 nucleotide posi- tions among members of this superfamily; however, most of these substitutions represent autapomorphies - POlMaS and do not provide strong support for any of the three possible relationships within this clade. Therefore, this

a a-73IM Tenthredlnidae

Tramax FIG. 4. (a) Parsimony tree of Hymenoptera based on 573 bp of nucleotide sequence from the 16s rRNA gene using a dipteran (Ae- des) as an outgroup. The minimum, maximum, and assigned num- bers of character state transformations are provided along each branch. This tree is 701 steps in length and has a CI of 0.753. (b) Bootstrap majority-rule consensus tree based on the same input ma- trix as the parsimony tree above. Numbers at each node represent Allorhogas the percentage of bootstrap trees supporting each branch.

Z’remex and the Aculeates, where only 57 and 46% of b the resampling procedures provided topologies sup- Tenthredinidae porting these relationships. Although this analysis shows clear relationships among the terebrants and Tremex among the aculeates, the basal branches of the tree, Apia corresponding to the relationships of the hymenop- Polistes teran suborders, are not strongly supported. XanthoplmpL, Current views of Hymenoptera classification support aax 73% a paraphyletic Symphyta basal to a monophyletic Apo- loa% Oigonogastra crita. Support (Fig. 4a) for a monophyletic clade that lea% Allorhogus includes the Symphyta and the aculeates may be due to the AT richness of these taxa relative to that of the Aphytts yanonansis remainder of the ingroup. Examination of less parsi- Aphytfs lingnanansis monious topologies revealed two trees of 702 steps that FIG. 5. (a) Parsimony tree of Hymenoptera based on 573 bp of differ from Fig. 4a only in partial resolutions of the nucleotide sequence from the 16s rRNA gene using a member of the ichnuemonoid clade. A single tree of 703 steps is found suborder Symphyta (Tenthredinidae) as an outgroup. The minimum, with the following relationships: (Tenthredinidae, maximum, and assigned numbers of character state transformations (Z’remex, (Aculeata, Terebrantes))); in other words, it are provided along each branch. This tree is 639 steps in length and has a CI of 0.79. (b) Bootstrap majority-rule consensus tree based on is congruent with current views. A strict consensus of the same input matrix as the parsimony tree above. Numbers at these four most parsimonious trees is substantially each node represent the percentage of bootstrap trees supporting less resolved than the tree presented in Fig. 4a. Tenth- each branch. mtDNA 16s rRNA SEQUENCE VARIATION FROM HYMENOPTERA 145 sequence seems inadequate for resolving relationships strated here, ribosomal sequence variation can provide among the ichneumonid Xunthopimplu and the two new insight into hymenopteran relationships; how- braconids. Conversely, the two chalcidoids, both in the ever, this sequence is clearly not a panacea. For ex- genus Aphytis, are clearly divergent and distinguished ample, problems associated with sequence alignment, on the tree by 47 and 60 character state transforma- primarily in regions with multiple gap events, compli- tions, Although their placement suggests that they are cates the phylogenetic analysis. In addition, strong bi- each other’s closest relative in our analysis, the ases in A and T bases render analyses restricted to amount and type of nucleotide variation expressed these characters less informative due to increased lev- from this genus were unexpected. This observation is els of nucleotide divergence between distantly related consistent with either very long divergence times, taxa. Finally, any examination across broad taxonomic which is not supported by their morphological similar- groups such as the Hymenoptera might well prove dif- ity, or phylogenetic sorting of ancestral mtDNA lin- ficult based on sequence divergence limited to one spe- eages among recently speciated taxa (Neigel and cific region. Nevertheless, studies such as the one pre- Avise, 1986). Nevertheless, it seems a more detailed sented here provide baseline information on the analysis of this gene region is warranted to better un- amount and type of nucleotide variation from specific derstand the relationships among the chalcidoids. regions and allow subsequent investigations on se- lected taxonomic groups to focus on areas of the ge- Utility of 16s rRNA Sequence for a Phylogenetic nome most likely to produce phylogenetically useful Analysis of Hymenoptera. We can now address the information. utility of this region for reconstructing hymenopteran Sequence Availability evolutionary history. First, 16s rRNA nucleotide se- quence variation may provide detailed information re- These sequences have been deposited in GenBank. garding the relationships of the aculeate Hymenop- tera. Although our sample was restricted to two members of this group, Apis (honeybee) and Polistes APPENDIX I (paper wasps), their sequence difference (d = 0.185) and the strong evidence for monophyly (Figs. 4 and Specimens Examined, Collection Localities, Method of 5) suggest that a more detailed examination at this Tissue Preservation, and Clone Numbers taxonomic level based on these nucleotide sequences Voucher specimens of all species have been deposited might be fruitful. This reasoning can also be applied as number 555 in the Insect Collection, Department of to the Chalcidoidea. Both representatives that we ex- Entomology, Texas A&M University. amined from this superfamily are members of the ge- nus Aphytis and the unexpected high sequence differ- Order Hymenoptera ence (d = 0.216) suggests further research with Suborder Symphyta additional representatives may provide a more de- Superfamily tailed understanding of the evolutionary history of this Siricidae group. In comparison, ichneumonoid sequences were Tremex columba (Linnaeusj-U.S.A.: Texas, remarkably conserved, showing little variation among Walker Co., 10 mi. W. New Waverly. Frozen the two families and three different subfamilies exam- specimen. TAMU Clones S-3-29 and S-15. ined. Based on these results, there seems little likeli- Superfamily hood of resolving relationships at the subfamily level Tenthredinidae-U.S.A.: Texas, Brazos Co., or above using these sequences. It will be necessary to Lick Creek Park. EtOH preserved specimen. sequence other members of the ephialtine Ichneumoni- Clone T4-36. dae, doryctine Braconidae, and braconine Braconidae Suborder Apocrita to determine if such sequences can be used to elucidate Superfamily Ichneumonoidea relationships among these genera. Conversely, high Ichneumonidae sequence divergence between members of the two hy- Xanthopimplu stemmutor (Thunberg)-U.S.A.: menopteran suborders results in instability at the base Texas, Brazos Co., College Station, lab culture. of the tree. Accordingly, nucleotide sequence informa- Frozen specimen. Clones C-7, C7-1, C7-12, tion from this 16s rRNA region is not informative in and C7-18. reconstructing the relationships among the suborders Braconidae of Hymenoptera. Digonogastra kimballi Kirkland-U.S.A.: Although nucleotide sequence variation holds great Texas, Brazos Co., College Station, lab cul- promise for enhancing our understanding of many tax- ture. Frozen specimen. Clones C5-21 and onomic groups that have proved intractable using tra- C5-22. ditional characters, the choice of an appropriate region Allorhogas pyralophagus Marsh-U.S.A.: for a given taxonomic level is paramount. As demon- Texas, Brazos Co., College Station, lab cul- 146 DERR ET AL

ture. Frozen specimen. Clones Cl-5, Cl-13, Gauld, I., and Bolton, B. (1988). “The Hymenoptera,” Oxford Univ. Cl-28, and C-41. Press, Oxford. Superfamily Chalcidoidea Gibson, G. A. P. (1986). Evidence for monophyly and relationships of Chalcidoidea, Mymaridae, and (Hymenoptera: Terebrantes). Can. Entomol. 118~ 205-240. Aphytis yanonensis DeBach and Rosen- Gibson, G. A. P. (1985). Some pro- and mesothoracic structures im- U.S.A.: Texas, Brazos Co., College Station, lab portant for phylogenetic analysis of Hymenoptera, with a review culture, originally from Shimizu-shi, Japan, of terms used for the structures. Can. Entomol. 117: 1395-1445. T89-145, frozen specimens from isolines. Harrison, R. G. (1989). Animal mitochondrial DNA as a genetic Clone C2-5. marker in populations and evolutionary biology. Trends Ecol. Aphytis lingnanensis Compere-U.S.A.: Euol. 4: 6-11. Texas, Brazos Co., College Station, lab Higgins, D. G., and Sharp, P. M. (1989). Fast and sensitive multiple culture, originally from Homestead, Florida, sequence alignments on a microcomputer. CABZOS 5: 151-153. Higgins, D. G., and Sharp, P. M. (1988). CLUSTAL: A package for T89-076. Clones ClO-23 and ClO-24. performing multiple sequence alignments on a microcomputer. Superfamily Gene 73: 237-244. Hillis, D. M., and Huelsenbeck, J. P. (1992). Signal, noise, and reli- Polistes versicolor-Venezuela: Maracay ability in molecular phylogenetic analyses. Hered., in press. nests larva. EtOH preserved specimen. Clones Hsu-Chen, C. C., Kotin, R. M., and Dubin, D. T. (1984). Sequence of C6-3 and C6-4. the coding and flanking regions of the large ribosomal subunit RNA gene of mosquito mitochondria. Nucleic Acids Res. 12: 7771-7785. Johnson, N. F. (1988). Midcoxal articulations and the phylogeny of ACKNOWLEDGMENTS the Order Hymenoptera. Ann. Entomol. Sot. Am. 81: 870-881. Kimura, M. (1980). A simple method for estimating evolutionary The authors thank J. W. Smith for supplying the specimens of rates of base substitutions through comparative studies of nucleo- Ichneumonoid taxa from his cultures, M. J. Rose for rearing and tide sequences. J. Mol. Euol. 16: 111-120. providing isolines of Aphytis, and T. M. Guerra, E. Arevalo, and D. Konigsmann, E. (1978a). Das phylogenetische System der Hymenop- Judd for laboratory assistance. J. W. Sites and B. Campbell provided tera. Teil 3: “Terebrantes” (Unterordnung Apocrita). Dtsch. Ento- helpful comments on the manuscript and J. M. Carpenter critically mol. 2. 25: l-55. evaluated and made substantial contributions to the sequence align- Konigsmann, E. (197813). Das phylogenetische System der Hymenop- ment and analysis sections. M. R. Thallman and J. F. Taylor pro- tera. Teil 4: Aculeata (Unterordnung Apocrita). Dtsch. Entomol. vided expert assistance with sequence analysis programs. This work Z., N. F. 25: 365-435. was supported by Texas Advanced Research Program Grant 999902- Konigsmann, E. (1977). Das phylogenetische System der Hymenop- 044 from the Texas Higher Education Coordinating Board and the tera. Teil 2: “Symphyta.” Dtsch. Entomol. 2. 24: l-40. Texas Agriculture Experiment Station. Konigsmann, E. (1976). Das phylogenetische System der Hymenop- tera. Teil I: Einfuehrung, Grundplanmerkmale, Schwestergruppe und Fossilfunde. Dtsch. Entomol. 2. 24: l-40. Kwiatowski, J., Skarecky, D., Hernandez, S., Pham, D., Quijas, F., REFERENCES and Ayala, F. J. (1991). High fidelity of polymerase chain reaction. Mol. Biol. Evol. 8: 884-887. Allard, M. W., Ellsworth, D. L., and Honeycutt, R. L. (1991). The Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). “Molecular production of single-stranded DNA suitable for sequencing using Cloning: A Laboratory Manual,” Cold Spring Harbor Laboratory the polymerase chain reaction. B&e&n&es 10: 24-26. Press, Cold Spring Habor, NY. Brothers, D. J. (1975). Phylogeny and classification of the aculeate McElroy, D., Moran, P., Bermingham, E., and Kornfield, I. (1991). Hymenoptera, with special reference to the Multillidae. Uniu. REAP: Restriction Enzyme Analysis Package, Version 4.0, pro- Kansas Sci. Bull. M): 483-648. gram manual, Department of Zoology, University of Maine, Orono. Carpenter, J. M. (1989). Testing scenarios: social behavior. Moritz, C., Dowling, T. E., and Brown, W. M. (1987). Evolution of CZu&stics 5: 131-144. animal mitochondrial DNA: Relevance for population biology and Carpenter, J. M. (1986). Cladistics of the . J. N. Y. systematics. Annu. Rev. Ecol. Syst. 18: 269-292. Entomd. Sot. 94: 303-330. Naumann, I. D., and Masner, L. (1985). Parasitic wasps ofthe procto- Clary, D. O., and Wolstenholme, D. R. (1985). The mitochondrial trupoid complex: A new family from Australia and a key to world DNA molecule of Drosophila yakuba: Nucleotide sequence, gene families (Hymenoptera: al.). Aust. J. Zool. 33: organization, and genetic code. J. Mol. Evol. 22 252-271. 761-783. Crozier, R. H., Crozier, Y. C., and Mackinlay, A. G. (1989). The CO-I Nei, M. (1987). “Molecular Evolutionary Genetics,” Columbia Univ. and CO-II region of honeybee mitochondrial DNA: Evidence for Press, New York. variation in insect mitochondrial evolutionary rates. Mol. Bid. Neigel, J. E., and Avise, J. C. (1986). Phylogenetic relationships of Evol. 8: 399-411. mitochondrial DNA under various demographic models of specia- Darling, D. C. (1988). Comparative morphology of the labrum in tion. In “Evolutionary Processes and Theory” (E. Nevo and S. Kar- Hymenoptera: the digitate labrum of and Euchariti- lin, Eds.), pp. 515-534, Academic Press, New York. dae (Chalcidoidea). Can. J. Zool. 68: 2811-2835. Rasnitsyn, A. P. (1988). An outline of evolution of the hymenopter- Eckert, K. A., and Kunkel, T. A. (1990). High fidelity DNA synthesis ous insects (Order Vespida). Orient. Insects 22: 115-145. by the Thermus aquaticus DNA polymerase. Nucleic Acids Res. Rasnitsyn, A. P. (1980). Origin and evolution of Hymenoptera. Tr. 1s: 3739-3744. Paleontol. Inst. 174: l-190. mtDNA 16s rRNA SEQUENCE VARIATION FROM HYMENOPTERA 147

Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., characterization of nucleic acid sequence functional domains. hiu- Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988). Primer- cleic Acids Res. 11: 2205-2220. directed enzymatic amplification of DNA with a thermostable Sokal, R. R., and Rohlf, F. J. (19811. “Biometry,” 2nd ed., Freeman, DNA polymeraae. Science 239: 487-491. San Francisco. Saini, M. S., and Dhillon, S. S. (1980). Metapleural transformations Swofford, D. L., and Olsen, G. J. (1990). Phylogeny reconstruction. with respect to the propodeum and metapostnotum in Hymenop- In “Molecular Systematics” (D. M. Hillis and C. Moritz, Eds.), pp. tera. Florida Entomol. 63: 286-291. 411-501, Sinauer Associates, Sunderland, MA. Sanger, F., Nicklon, S., and Coulson, A. R. (1977). DNA sequencing Swofford, D. L. (1985). PAUP: Phylogenetic Analysis Using Parsi- with chain terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: mony, Version 3.0, program manual, Illinois Natural History Sur- 5463-5461. vey, Urbana. Simon, C., Pairbo, S., Kocher, T. D., and Wilson, A. C. (1990). Evolu- Vlasak, I., Burgschwaiger, S., and Kreil, G. (1987). Nucleotide se- tion of the mitochondrial ribosomal RNA in insects 88 shown by quence of the large ribosomal RNA of honeybee mitochondria. Nu- the polymerase chain reaction. In “Molecular Evolution” (M. T. cleic Acids Res. 16: 2388. Clegg and S. J. O’Brien, Eds.), New Series, Vol. 122, pp. 235-244, Whitfield, J. B., Johnson, N. F., and Hamerski, M. R. (1989). Identity UCLA Symposia on Molecular and Cellular Biology. and phylogenetic significance of the metapostnotum in nonacule- Smith, T. F., Waterman, M. S., and Sadler, J. R. (1983). Statistical ate Hymenoptera. Ann. Entomol. Sot. Am. 82: 663-673.